Template recognition by an RNA-dependent RNA polymerase: identification and characterization of two RNA binding sites on Q beta replicase

Two different SELEX protocols were used to generate two classes of RNA ligands that bound Q beta replicase with nanomolar equilibrium dissociation constants. One set of RNAs appeared to exist as pseudoknots with conserved loop sequences. These ligands bound Q beta replicase and ribosomal protein S1 with equal affinities, indicating that the RNAs bind the replicase through its S1 subunit. The second class of ligands bound the replicase via a pyrimidine rich region. The two sets of ligands did not compete for binding to Q beta replicase, demonstrating that the two RNA families bind unique sites on the phage polymerase. Because the second class of ligands bound ribosomal protein S1 very poorly, it is likely that the second RNA binding site is located on one of the three remaining replicase subunits. Published sequences of RNAs replicated by Q beta replicase possess similarities to the two classes of RNA ligands, providing a possible solution to the question of template recognition by the phage polymerase.

Selection and characterization of RNAs replicated by Q beta replicase

RNAs replicated by Q beta replicase were isolated from two random sequence RNA populations (one 56 nucleotides in length, the second 83) using a replication/dilution protocol. The selected molecules were cloned and sequenced, generating data set of 54 replicatable RNAs bound with higher affinity to Q beta replicase than did the random populations from which they were selected. Deletion analyses on two of the molecules indicated that internal regions of the RNAs were responsible for the specific binding of Q beta replicase. Truncated molecules representing the minimized RNA binding sites could inhibit replication of the full-length molecules, apparently by obstructing their binding to the replicase. The binding regions of the two RNAs were dominated by extended runs of pyrimidines. Similar C/U-rich regions existed in 85% of the sequences in the data set as well as in all of the previously published replicatable sequences. Mutation of the polypyrimidine domain of one of the replicatable sequences reduced the affinity of the molecule for Q beta replicase by 10-fold and completely abolished its ability to be replicated.

Let's get specific: the relationship between specificity and affinity

The factors that lead to high-affinity binding are a good fit between the surfaces of the two molecules in their ground state and charge complementarity. Exactly the same factors give high specificity for a target. We argue that selection for high-affinity binding automatically leads to highly specific binding. This principle can be used to simplify screening approaches aimed at generating useful drugs.

RNA ligands to human nerve growth factor

High affinity RNA ligands to human nerve growth factor (NGF) were selected from pools of random RNA using SELEX [Tuerk, C. and Gold, L. (1990) Science, 249, 505-510]. Nerve growth factor, which is a protein required for the development of neurons, is not known to bind nucleic acids as part of its natural function. We describe two of the selected RNA molecules in detail. One of them is highly structured, folding into a pseudoknot with an additional hairpin-loop; this structure provides salt-resistant binding to NGF. The other is unstructured and elevated salt concentrations inhibit its binding. These molecules compete with each other for NGF binding. Our RNAs may furnish useful diagnostic tools for the study of an important neurotrophic protein; additionally, they illustrate another example of the potential for nucleic acids to take part in novel binding interactions.

High-affinity ssDNA inhibitors of the reverse transcriptase of type 1 human immunodeficiency virus

The reverse transcriptase (RT) of HIV-1 is a plausible target for therapeutic agents aimed at inhibiting propagation of the virus. We have used "irrational drug design", that is, combinatorial chemistry with oligonucleotide libraries, to identify high-affinity ligands aimed at HIV-1 RT. The methodology, termed SELEX (systematic evolution of ligands by exponential enrichment), was employed with a single-stranded DNA library. The selected ssDNA ligands bind HIV-1 RT with Kd values as low as 1 nM and inhibit the RNA-dependent DNA-polymerase activity of the enzyme with Ki values as low as 0.3 nM. We also demonstrate the high specificity of one ligand able to selectively discriminate between the reverse transcriptases of HIV-1, AMV, and MMLV. These ssDNA molecules may be useful as inhibitors or as models for the design of small molecule inhibitors of HIV-1 RT in vivo.

Isolation of high-affinity RNA ligands to HIV-1 integrase from a random pool

We were able to isolate high-affinity RNAs from a random pool that binds to integrase protein from the human immunodeficiency virus-type 1 using the procedure now known as SELEX. Generally, the RNAs fell into three different classes in binding buffer containing 250 mM NaCl: group I class of molecules binds integrase with a dissociation constant (Kd) on the order of 10 nM, group II molecules had a Kd of about 80 nM, and group III about 800 nM. The RNA with the highest affinity from the group I class of molecules, designated P5, was characterized using computer modeling, chemical and enzymatic probing, and deletion analysis. Our secondary structure model for this RNA suggests interactions between looped-out fixed nucleotides and nucleotides from the randomized region; a GNRA tetraloop is also in the structure. We showed that our integrase was able to process a U5 mimic in vitro. P5 competes effectively for binding with the double-stranded DNA mimic of U5 at 180 mM NaCl concentration.

Diversity of oligonucleotide functions

SELEX is a technology for the identification of high affinity oligonucleotide ligands. Large libraries of random sequence single-stranded oligonucleotides, whether RNA or DNA, can be thought of conformationally not as short strings but rather as sequence dependent folded structures with high degrees of molecular rigidity in solution. This conformational complexity means that such a library is a source of high affinity ligands for a surprising variety of molecular targets, including nucleic acid binding proteins such as polymerases and transcription factors, non-nucleic acid binding proteins such as cytokins and growth factors, as well as small organic molecules such as ATP and theophylline. The range of applications of this technology for new discovery extends from basic research reagents to the identification of novel diagnostic and therapeutic reagents. Examples of these applications are described along with a discussion of underlying principles and future developments expected to further the utility of SELEX.

Crystal structure of the T4 regA translational regulator protein at 1.9 A resolution

The translational regulator protein regA is encoded by the T4 bacteriophage and binds to a region of messenger RNA (mRNA) that includes the initiator codon. RegA is unusual in that it represses the translation of about 35 early T4 mRNAs but does not affect nearly 200 other mRNAs. The crystal structure of regA was determined at 1.9 A resolution; the protein was shown to have an alpha-helical core and two regions with antiparallel beta sheets. One of these beta sheets has four antiparallel strands and has some sequence homology to RNP-1 and RNP-2, which are believed to be RNA-binding motifs and are found in a number of known RNA-binding proteins. Structurally guided mutants may help to uncover the basis for this variety of RNA interaction.

Active macrophage-associated TGF-beta co-localizes with type I procollagen gene expression in atherosclerotic human pulmonary arteries

Vascular remodeling in adult atherosclerotic pulmonary arteries is characterized by discrete areas of neointimal smooth muscle cell extracellular matrix gene expression in close proximity to non-foamy macrophages, suggesting regulation by local macrophage-associated factors. The purpose of these studies was to begin addressing the role of putative macrophage-associated factors such as transforming growth factor-beta (TGF-beta), by determining the spatial relationship between TGF-beta and neointimal matrix gene expression in human atherosclerotic pulmonary arteries. For example, the participation of TGF-beta in vascular remodeling could be inferred by its colocalization with non-foamy macrophages in areas of active matrix synthesis. In situ hybridization and immunohistochemistry demonstrated focal neointimal procollagen gene expression in close association with non-foamy but not foamy macrophages. Immunohistochemistry with isoform-specific anti-TGF-beta antibodies demonstrated all three isoforms of TGF-beta associated with non-foamy macrophages, but foamy macrophages were not immunoreactive. Neointimal and medial smooth muscle cells stained lightly. In contrast, intense TGF-beta immunoreactivity was also associated with medial smooth muscle cells in normal nonremodeling vessels. Immunohistochemistry with antibodies specific for latent TGF-beta was similar to immunohistochemistry for mature TGF-beta in both remodeling and nonremodeling vessels. Finally, using an antibody specific for active TGF-beta 1, immunoreactivity was only seen in non-foamy neointimal macrophages but not in foamy macrophages or medial smooth muscle cells from hypertensive or normal vessels. These observations suggest non-foamy macrophages may participate in modulating matrix gene expression in atherosclerotic remodeling via a TGF-beta-dependent mechanism.

In vitro selection of RNA ligands to substance P

RNA ligands to the tachykinin substance P have been selected from a large pool of random sequence RNA molecules. Substance P is an undecapeptide that plays a variety of roles as a neurotransmitter and neuromodulator in the central and peripheral nervous system of mammals. A systematic evolution of ligands by exponential enrichment (SELEX) procedure was used to isolate RNAs that bind substance P immobilized on a solid support. RNAs that also bind substance P in solution were identified, and the tightest binder was subjected to mutagenesis in a second SELEX procedure to evolve ligands with a higher affinity for the peptide. A comparative analysis of 36 ligands isolated from the second SELEX experiment revealed two main sequence classes with highly conserved secondary structures within each class. Dissociation constants for the interaction of these ligands with substance P in solution were determined by equilibrium dialysis. The amino acid residues involved in the interaction with the highest affinity ligand (190 nM Kd) were mapped by determining which of a set of overlapping fragments of substance P can compete with the intact peptide for binding. A binding competition experiment also demonstrated the ability of the same ligand to discriminate between substance P and the reverse orientation of the same amino acid sequence. The results from this study demonstrate that SELEX can yield high affinity RNA ligands to small nonconstrained peptides.

High-affinity RNA ligands to Escherichia coli ribosomes and ribosomal protein S1: comparison of natural and unnatural binding sites

High-affinity RNA ligands were generated against intact 30S ribosomes, S1-depleted 30S ribosomes, and purified ribosomal protein S1. Sequence analysis indicated two classes of ligand: unstructured RNAs containing a Shine-Dalgarno sequence and structured RNAs containing a pseudoknot. The Shine-Dalgarno-containing ligands were generated against S1-depleted 30S ribosomes but, surprisingly, not against intact 30S ribosomes or ribosomal protein S1. In contrast, pseudoknot-containing ligands were generated against intact ribosomes as well as purified S1 protein. The two classes of ligand exhibited specificity for their respective targets, as well as conserved sequence and secondary structure reminiscent of naturally occurring, cis-acting mRNA elements.

Conformational properties of oligonucleotides

Vast libraries of oligonucleotide sequences can be made. Such libraries have been screened and/or selected to identify individual molecules showing various properties, including high affinity/high specificity binding (to something) or new enzymatic activity. The SELEX protocol is used to identify these interesting oligonucleotides; two recent reviews of the SELEX protocol and its power have appeared recently. Oligonucleotides derived through the SELEX process have applications in diagnostics and therapeutics, and have additional value as research reagents and compounds that provide chemical plausibility to some ideas about evolution of the biosphere. Oligonucleotide conformations are shockingly robust and potent.

Changes in CNS neuropeptide FF-like material, pain sensitivity, and opiate dependence following chronic morphine treatment

Tolerance and dependence to opiates may be an adaptive process that limits excessive effects of morphine on the CNS. Because no consistent opiate receptor reduction in chronically treated rats seems to underlie the hyposensitivity to morphine, an alternative hypothesis has postulated a role of "antiopioid" peptides. It is possible to speculate that the administration of morphine stimulates antiopioid systems such as neuropeptide FF (NPFF), as part of an homeostatic mechanism contributing to the development of tolerance. To test this hypothesis, pain sensitivity, opiate dependence, and CNS NPFF-IR levels were estimated at different times after implantation of morphine pellets (2 x 75 mg; NIDA). Three hours after morphine pellet treatment the analgesic effect was maximum and it decreased rapidly during the following 12 h. Naloxone-precipitated withdrawal syndrome was detected as soon as 3 h after morphine pellet implantation and was maximal after 24 h. NPFF-IR levels were measured in the spinal cord, brain stem, and hypothalamus. A significant decrease of NPFF-IR was observed 1 h after morphine pellet implantation (-25% to -45% depending on the structures) followed by a drastic increase of NPFF-IR levels (+60 to +140%) between 3 and 6 h. NPFF-IR levels rapidly returned to baseline after 24-36 h. It is suggested that the activity of these NPFF-IR neurones may increase gradually as a consequence of the continuous stimulation of opiate receptors and be part of an adaptive process that is able to counteract morphine effects and to induce dependence and tolerance to the analgesic effects of opiates.

Enrichment for RNA molecules that bind a Diels-Alder transition state analog

RNA molecules that bind a transition state analog for a Diels-Alder reaction (Kd = 0.35 +/- 0.05 mM) were isolated from a starting pool of approximately 10(14) sequences by affinity chromatography. After the initial rise and plateau of the amount of RNA that eluted with soluble analog, a step gradient elution was used to further enrich the pool for sequences with higher affinities for the target. To our knowledge, the isolation of RNA molecules that bind either a nonplanar or a hydrophobic ligand has not been reported previously. A conserved nucleotide sequence and secondary structure present in many of the RNA molecules are necessary but not sufficient for binding the analog. No catalysts of the targeted Diels-Alder reaction were found among the binders. The absence of catalysis contrasts with previous successful experiments with antibodies and suggests that other strategies may be needed to identify oligonucleotides with diverse catalytic activities.

The bacteriophage T4 regB ribonuclease. Stimulation of the purified enzyme by ribosomal protein S1

Infection of Escherichia coli by bacteriophage T4 induces a mRNA ribonuclease activity that shows specificity for cleavage within the sequence GGAG. Substrates of the activity in vivo include a number of phage mRNAs that are cleaved at GGAGs within the Shine/Dalgarno domains of their translation initiation regions. Induction of the ribonuclease depends on the product of the T4 gene regB. We describe here the overproduction and extensive purification of the RegB protein. RegB precisely co-purifies with an activity that cleaves within the sequence GGAG in oligonucleotide and polynucleotide RNAs and is therefore likely to constitute the sequence-specific catalytic component of the observed activity. We further report that the low cleavage rate observed with our preparations of purified RegB is substantially increased (1-2 orders of magnitude) by the addition of E. coli ribosomal protein S1. We discuss the implications of this observation for the mechanism of action of the RegB ribonuclease in vitro and in vivo.

Selection of high-affinity RNA ligands to reverse transcriptase: inhibition of cDNA synthesis and RNase H activity

Specific, high-affinity RNA ligands to avian myeloblastosis virus and Moloney murine leukemia virus reverse transcriptases were isolated from a combinatorial RNA library using the SELEX (systematic evolution of ligands by exponential enrichment) procedure. The selected RNA ligands bound their respective reverse transcriptases with approximately nanomolar dissociation constants. The ligands did not exhibit primary sequence conservation from selections against different target enzymes. Moreover, the selected ligands competed with the binding of template/primer complex and inhibited both the RNA-dependent DNA polymerase and the RNase H activities of the cognate reverse transcriptase. SELEX can yield both high-affinity and high-specificity oligonucleotide antagonists against specific members of a protein family.

TGF beta 1 and TGF beta 2 are potential growth regulators for medulloblastomas, primitive neuroectodermal tumors, and ependymomas: evidence in support of an autocrine hypothesis

Our previous investigations of transforming growth factor types beta 1 and beta 2 (TGF beta s) showed negative or positive autocrine growth regulation of gliomas in vitro. Near-diploid gliomas were inhibited by the TGF beta s, whereas a stimulatory response correlated with progressive anaplasia and karyotypic divergence. We have tested the hypothesis that cytogenetic aberrations may be associated with conversion of TGF beta autoregulation from inhibitory to stimulatory among other cultured neuroectodermal tumors. Anchorage-independent growth and karyotypic aberrations supported the malignant nature in vitro of two medulloblastoma (MBL), two primitive neuroectodermal tumor (PNET), and two ependymoma (EPD) cultures. Transforming growth factor type beta 1 and/or TGF beta 2 RNA was evident by Northern blot analysis among these cell cultures. By radioreceptor assay active TGF beta was present in conditioned medium in concentrations of 0 to 14 ng/mL, whereas the total amount of active and latent TGF beta secreted was in the range of 3 to 118 ng/mL. Expression of the TGF beta radioreceptor (TGF beta-R) types I and II was shown by cross-linking assay. Responses to exogenous TGF beta were determined by [3H]-thymidine incorporation, cell counts, and anchorage-independent clonogenicity. Exogenous TGF beta was growth inhibitory for the near-diploid MBL, PNET, and EPD in vitro, as well as antagonistic to the mitogenic effect of epidermal growth factor (EGF) and insulin. In contrast, MBL, PNET, and EPD with a hyperdiploid subpopulation were stimulated to proliferate in monolayer culture or soft agar by TGF beta 1 and TGF beta 2. The growth response did not correlate with TGF beta-R type. Autocrine regulation was supported by antibody neutralization experiments performed with quiescent cells in the absence of exogenous TGF beta. Anti-TGF beta antisera enhanced the growth of TGF beta-inhibited cultures, whereas the TGF beta-stimulated cultures were inhibited by the antisera. Karyotypic divergence seemed to predict response as MBL, PNET, and EPD with hyperdiploid elements exhibited autocrine TGF beta-stimulation. In contrast, the near-diploid cultures were inhibited by the TGF beta s. By analogy with the gliomas, conversion of TGF beta autocrine regulation from inhibition to stimulation may be a late progression marker of anaplasia among MBL, PNET, and EPD. Secretion of this TGF, which serves both as a mitogen and immunosuppressive agent may contribute to the adverse prognosis of hyperdiploid neuroectodermal neoplasms of the central nervous system (CNS).

Quantitative analysis of ribosome binding sites in E.coli

185 clones with randomized ribosome binding sites, from position -11 to 0 preceding the coding region of beta-galactosidase, were selected and sequenced. The translational yield of each clone was determined; they varied by more than 3000-fold. Multiple linear regression analysis was used to determine the contribution to translation initiation activity of each base at each position. Features known to be important for translation initiation, such as the initiation codon, the Shine/Dalgarno sequence, the identity of the base at position -3 and the occurrence of alternative ATGs, are all found to be important quantitatively for activity. No other features are found to be of general significance, although the effects of secondary structure can be seen as outliers. A comparison to a large number of natural E.coli translation initiation sites shows the information profile to be qualitatively similar although differing quantitatively. This is probably due to the selection for good translation initiation sites in the natural set compared to the low average activity of the randomized set.

Recognition of the mRNA selenocysteine insertion sequence by the specialized translational elongation factor SELB

In Escherichia coli the unusual amino acid selenocysteine is incorporated cotranslationally at an in-frame UGA codon. Incorporation of selenocysteine relies, in part, on the interaction between a specialized elongation factor, the SELB protein, and a cis-acting element within the mRNA. Boundary and toeprint experiments illustrate that the SELB-GTP-Sec-tRNA(Sec) ternary complex binds to the selenoprotein encoding mRNAs fdhF and fdnG, serving to increase the concentration of SELB and Sec-tRNA(Sec) on these mRNAs in vivo. Moreover, toeprint experiments indicate that SELB recognizes the ribosome-bound message and that, upon binding, SELB may protrude out of the ribosomal-mRNA track so as to approach the large ribosomal subunit. The results place the mRNA-bound SELB-GTP-Sec-tRNA(Sec) ternary complex at the selenocysteine codon (as expected) and suggest a mechanism to explain the specificity of selenocysteine insertion. Cis-acting mRNA regulatory elements can tether protein factors to the translation complex during protein synthesis.

Direct transfer of transforming growth factor beta 1 gene into arteries stimulates fibrocellular hyperplasia

The arterial wall responds to thrombosis or mechanical injury through the induction of specific gene products that increase cellular proliferation and connective tissue formation. These changes result in intimal hyperplasia that is observed in restenosis and the early phases of atherosclerosis. Transforming growth factor beta 1 (TGF-beta 1) is a secreted multi-functional protein that plays an important role in embryonal development and in repair following tissue injury. However, the function of TGF-beta 1 in vascular cell growth in vivo has not been defined. In this report, we have evaluated the role of TGF-beta 1 in the pathophysiology of intimal and medial hyperplasia by gene transfer of an expression plasmid encoding active TGF-beta 1 into porcine arteries. Expression of TGF-beta 1 in normal arteries resulted in substantial extracellular matrix production accompanied by intimal and medial hyperplasia. Increased procollagen, collagen, and proteoglycan synthesis in the neointima was demonstrated by immunohistochemistry relative to control transfected arteries. Expression of TGF-beta 1 induced a distinctly different program of gene expression and biologic response from the platelet-derived growth factor B (PDGF B) gene: procollagen synthesis induced by TGF-beta 1 was greater, and cellular proliferation was less prominent. These findings show that TGF-beta 1 differentially modulates extracellular matrix production and cellular proliferation in the arterial wall in vivo and could play a reparative role in the response to arterial injury.

Translation initiation complex formation with 30 S ribosomal particles mutated at conserved positions in the 3'-minor domain of 16 S RNA

Escherichia coli 30 S ribosomal subunits containing in vitro (phage T7 RNA polymerase-generated) 16 S rRNA, both wild-type and mutant, were examined by toeprinting. These synthetic particles were used to compare the effects of the absence of base modification and of specific nucleotide substitutions in conserved sequence regions of the RNA on the assembly of mRNA, tRNAs and 30 S particles into a translational initiation complex. Initiation factor-3-dependent selection of tRNA(fMet) from a mixture of tRNA(fMet) and tRNA(Phe) occurred with all particles, although 20 times less initiation factor-3 was needed for the synthetic particles, including the mutants. Whereas isolated 30 S particles or those reconstituted with isolated RNA did not distinguish between tRNA(fMet) and tRNA(Phe) for ternary complex formation in the absence of initiation factor-3 (intrinsic selection ability), the synthetic particles preferred tRNA(fMet). The difference between the natural and synthetic particles appears to be due to the absence of certain base modifications, but not m2(6)A, in the synthetic RNA. Synthetic particles containing the mutation U1512C, which converts the universal U.G pair to C.G enhanced both tRNA(fMet) binding and selectivity, although other mutations at that site, namely U1512G, G1523A and U1512C/C1524U, had no such effect. Mutants U1498G and G1401C/C1501G, both located in a highly conserved single-stranded region of the 3'-minor domain, also enhanced tRNA(fMet) selectivity, in this case by reducing complex formation with elongator tRNA. Complex formation between elongator tRNA and the G1401C/C1501G mutant was reduced to almost undetectable levels. The results also indicated that the association rate for initiation complex formation for G1401C/C1501G was considerably lower than for the wild-type sequence. This result had not been detected by standard tRNA-30 S binding assays. Overall, the data suggest that (some of) the 16 S rRNA base modifications as well as the tertiary structure around the decoding site act to desensitize the intrinsic selection ability of the ribosome for tRNA(fMet).